MECH-VISION PRO Visual Data & Key Point Training
Spearheaded the data structuring, visual processing, and model training pipeline for an advanced diagnostic tool utilizing MediaPipe and TensorFlow Lite. This project required strict attention to detail in processing spatial data and mapping visual inputs to ensure the AI could accurately interpret real-time conditions. Responsibilities included validating key point detection, testing the accuracy of the computer vision model against complex mechanical structures, and refining the dataset to eliminate false positives. The rigorous data handling directly improved the model's diagnostic reliability and processing efficiency.